Ioannis CanellosHead of Software Engineering at Cyber Ranges. Member of the Apache Software Foundation. 20 years building software, 15 in open source.
Creator of Fabric8 Kubernetes Client, Spring Cloud Kubernetes, Dekorate, Sundrio, and Apache Karaf Cellar.
Previously Senior Principal Software Engineer at Red Hat on the Quarkus team (2012–2025).
In my previous post Workflows everywhere pt. 1 I tried to define workflows and enumerate their functional and non-functional requirements. The post concluded with the realization that in many case we need workflow engines to power our workflows.
This post defines what a workflow engine is and lists some of the most popular engines by category. Or at least that was the original intention, but there is a twist.
Workflow engines are systems designed to simplify the creation and execution of workflows. They orchestrate the flow of information between the activities that compose the workflow based on predefined logic, conditions and dependencies.
Workflows are everywhere. From CI/CD pipelines, all system / data integration to business process automation. It wouldn’t be too far-fetched to say that even modern software build tools like make, maven or npm are in their own way workflow engines.
There are countless tools out there that help people define, execute and monitor workflows varying from simple no-code tools to complex frameworks that allow developers to define workflows in code, or even architect their software as workflows.
I recently had a discussion with two fellow engineers about secure coding practices. After the discussion I realized that I am neglecting one of the most important practices: signing my commits.
There are tons of articles on the internet explaining why and how. These are my notes on the subject that I decided to publish.
These notes actually use literate programming so they are a mix of notes and code you can actually use via org-mode.
You can find the actual notes file here.
It seems that everyone is an MCP guru these days.
I am not.
In fact, I know almost nothing about it. I am just aware of the concept.
This post describes the steps I took in order to create an MCP from scratch, resulting in the project created at:
I also recorded my journey starting from scatch with almost zero knowledge on the topic and using the Quarkus Blog as a guide:
Back in April 2020 I shared how I built a nutrition tracker in Emacs that leveraged org-capture templates and or-ql to record foods, recipes, and meals. At that time, I relied on an org-mode based database and manual updates to keep track of calories, protein, carbs, and fat. While the system worked, maintaining that data was both tedious and error-prone. Each time I needed to insert a new food, I had to do an internet search to find the nutritional information and then manually update my org-mode files.
A couple of weeks ago I came across Roberto Carratalla’s blog post on Function calling on OpenShift AI. At the time I was preparing for RedHat Summit Connect Zurich 2025 where I was meant to run a workshop on Quarkus and Langchain4j with Kyra Goud, Dimitris Andreadis and Codrin Bucur. We had an issue however, related to enabling functions on OpenShift AI.
Roberto pointed us to the blog post, but I couldn’t spot what I was doing wrong. So, I decided to port the examples in the blog post to Java to make sure that I was comparing apples to apples.